trueflu :
This is my point. I didn't address the cost. I've noted specific examples of the benefits of optimization. You just say I'm not right, not why I'm not right.
This is my point, you don't iclude cost in both time and money, and you cannot speak intelligently to this subject.
You seem to think if Crytek just sent a million bucks to some account somewhere then Crysis would be more optimized.
Most games are optimized, within the time frame of their development cycle, and even then they still ship late almost every time.
Spend too much time optimizing for this years hardware/API by the time you ship new hardware and API are out that make your actions maybe not worthless, but worth less.
With a robust market the cost, even if so awful, would be worth it.
Based on what other than your guess?
I know CUDA won't help the gaming market, but it is a very pertinent example of Optimizing in the real world. I didn't claim we'd get more gaming power out of it.
CUDA's not about optimizing, it's about utilizing, it a tool that allows you different options,. Optimizing would be what is done with CUDA or CTM to achieve the task as efficiently as possible. Your stateent's like saying they shouldn't use OpenGL they should use DX to optimize games, or the inverse of that, as if either would solve the problem of 'optimization'.
Macs have far better optimizing on their OS's than do PC's. Again a pertinent example of what optimizing can do. Not an example of how GPU optimizing should be applied.
Not an example at all, more of an assumption on your part, and one without support.
OSX does some things more effificiently and others less so. Once again it's like you invoking the term CUDA with little to no understanding of how it's supposed to help other than reading somewhere that "cuda programming mutiplies the power of gpu's a hundred fold" as if it were just like adding a part to the computer as an overclocker.
I think Spongebob's right, your lack of understanding is going to mean that you're going to complain about this abstract idea of 'optimization' without the ability to even understand the answer.
Simply put for you, Optimization takes time and money, the return on that investment needs to be large enough to warrant it, mostly it's not since no one cares if the low end doesn't have all the options as long as the high end buyers do since they're buying on launch day while the bargain gamer is waiting for the price drop.
Also, it's better to build it efficiently from the start that to try and fix it at the end with optimizations. The biggest problem for that with cutting edge games and hardware is that the platforms/hardware you're building your game to play on don't exist when you're in early-mid development. Your request to build for the middle means you push the game development timeframe back a year or two. So either you Can have Crysis last fall or next year. Most people would take last fall, and learn to Move The Sliders To The LEFT.
(edited for typos after looking at the quote below)